๐Ÿ‡ง๐Ÿ‡ช Based in Belgium ยท EU-Compliant ยท GDPR-Ready

Building the Future
with Intelligent AI

A next-generation AI development company based in Belgium, specialising in Large Language Models, Small Language Models, and enterprise AI solutions โ€” built on the world's best cloud infrastructure.

LLM + SLM
Model Portfolio
3 Live
AI Projects
4 Clouds
Multi-Cloud Strategy
๐Ÿ‡ช๐Ÿ‡บ EU
Data Sovereignty

Our Core Services

End-to-end AI development, from model selection to production deployment.

๐Ÿง 

LLM & SLM Integration

We integrate state-of-the-art large and small language models into your existing enterprise systems. Fine-tuning, RAG pipelines, and custom model orchestration included.

๐Ÿค–

AI Agent Development

Build autonomous AI agents for task automation, customer support, document processing, and decision support using agentic frameworks like LangGraph and AutoGen.

๐Ÿ”

RAG & Knowledge Systems

Enterprise-grade Retrieval-Augmented Generation (RAG) systems that connect your private data to AI models โ€” securely and with full EU data residency.

โ˜๏ธ

Multi-Cloud AI Infra

Architecture and deployment on AWS, Google Cloud, Azure, and OVHcloud. We optimise cost and design for scalability from day one.

๐Ÿ“Š

AI POC & Prototyping

Rapid proof-of-concept builds that validate AI use cases before full investment. Typical POC timelines: 2โ€“6 weeks with clear success metrics.

๐Ÿ”’

Compliant AI Solutions

GDPR-compliant, EU AI Act-aligned solutions built with data privacy at the core. Particularly suited for regulated industries in Belgium and the EU.

Featured Projects

Real AI systems we've architected and deployed โ€” from on-premise Ollama agents to cloud-native RAG platforms.

ollama_mcp_agent.py โ€” OllamaDesk Core
1# OllamaDesk ยท Local AI Assistant via Ollama + MCP
2from ollama import AsyncClient
3from mcp import MCPToolServer, Tool
4from langgraph.graph import StateGraph
5 
6# Bind local Llama 4 Scout via Ollama
7llm = AsyncClient(model="llama4:scout")
8 
9# Register MCP Tool Servers
10mcp = MCPToolServer()
11mcp.register("filesystem")  # read/write files
12mcp.register("calendar")    # MS365 calendar
13mcp.register("postgres_db")  # SQL queries
14 
15# LangGraph agent loop
16graph = StateGraph(AgentState)
17graph.add_node("reason", llm_reason_step)
18graph.add_node("act", mcp.execute_tool)
19graph.add_edge("reason", "act")
20 
21async def run_agent(query: str) -> str:
22    return await graph.ainvoke({"input":query})
Project 01
โ— Live

OllamaDesk โ€” On-Premise AI Desk Assistant

Local LLMs + Model Context Protocol (MCP) = Zero data exposure

OllamaDesk is a fully on-premise enterprise AI assistant powered by Ollama running Llama 4 Scout locally, connected to enterprise tools via Anthropic's Model Context Protocol (MCP). No data ever leaves the corporate network โ€” making it ideal for GDPR-sensitive Belgian enterprises. A LangGraph ReAct loop orchestrates the reasoning and tool-execution cycle.

๐Ÿฆ™

Ollama Local Inference

Runs Llama 4 Scout, Mistral Nemo, and Phi-4 entirely on-premise โ€” GPU or CPU inference, no API keys needed.

๐Ÿ”Œ

MCP Tool Integration

MCP servers expose filesystem, Microsoft 365 Calendar, PostgreSQL, and SharePoint as native tools the AI can invoke.

๐Ÿ”„

LangGraph ReAct Agent Loop

Stateful multi-step reasoning: the agent plans, calls tools, observes results, and self-corrects until the task completes.

Ollama MCP (Anthropic) LangGraph Llama 4 Scout Phi-4 FastAPI Docker
< 200ms
First Token
0 bytes
Data Egress
8 tools
MCP Servers
3 models
Switchable
Project 02
โ— Live

DocuFlow AI โ€” Multilingual Document Intelligence

RAG pipeline for NL / FR / EN enterprise documents

DocuFlow AI is a production-grade Retrieval-Augmented Generation (RAG) platform built for Belgian enterprises that need to query large internal document repositories in Dutch, French, and English simultaneously. It uses Mistral Nemo for EU-native multilingual embeddings and Command R+ for grounded, citation-backed answers โ€” all deployed on OVHcloud for GDPR compliance.

๐Ÿ—‚๏ธ

Hybrid Search (Dense + BM25)

Weaviate vector store with hybrid retrieval combining dense semantic search and keyword BM25 for maximum recall across document types.

๐ŸŒ

Trilingual NLP (NL / FR / EN)

Automatic language detection and cross-lingual retrieval โ€” ask in French, retrieve Dutch documents, get a unified English answer.

๐Ÿ“Œ

Grounded Citations

Every response is backed by direct source citations with document name, page number, and confidence score โ€” no hallucinations.

Mistral Nemo Command R+ Weaviate LangChain OVHcloud FastAPI React
3 langs
NL/FR/EN
1M+ docs
Indexed
94%
Accuracy
EU only
Data Region
๐Ÿ” DocuFlow AI โ€” RAG Pipeline Architecture
๐Ÿ“„

Document Ingestion

PDF, DOCX, PPTX, SharePoint, OneDrive โ€” auto-parsed & chunked

โ†“ chunking + lang detection
๐Ÿงฌ

Multilingual Embedding

Mistral Nemo embeddings ยท 1024-dim vectors for NL/FR/EN

โ†“ upsert to vector store
๐Ÿ—„๏ธ

Weaviate Hybrid Store

Dense vector + BM25 keyword indices on OVHcloud EU

โ†“ hybrid retrieval + rerank
๐Ÿค–

Command R+ Generator

Grounded answer synthesis with inline source citations

โ†“ structured response

๐Ÿ’ฌ Chat UI

React frontend

โšก REST API

FastAPI endpoints

smartops_agent.py โ€” Incident AutoRemediation
1# SmartOps Agent ยท IT Incident Management with AI
2from autogen import ConversableAgent, GroupChat
3from mcp import MCPToolServer
4import anthropic
5 
6# Tool integrations via MCP
7ops_tools = MCPToolServer()
8ops_tools.register("servicenow") # ITSM tickets
9ops_tools.register("prometheus") # metrics
10ops_tools.register("kubectl")      # k8s control
11ops_tools.register("pagerduty")    # alerting
12 
13# Multi-agent group: Analyst + Executor + Reviewer
14analyst = ConversableAgent("analyst", llm="claude-sonnet-4-5")
15executor = ConversableAgent("executor", tools=ops_tools)
16reviewer = ConversableAgent("reviewer", llm="claude-sonnet-4-5")
17 
18# Trigger: P1 alert โ†’ autonomous resolution
19chat = GroupChat([analyst, executor, reviewer])
20await chat.run(incident_payload)
Project 03
โ— Beta

SmartOps โ€” Autonomous IT Incident Agent

Multi-agent AI that detects, diagnoses, and auto-remediates IT incidents

SmartOps is a multi-agent AI system powered by Microsoft AutoGen and Claude Sonnet 4.5, connected to enterprise IT tools via MCP. When a P1/P2 alert fires from Prometheus or PagerDuty, three specialised AI agents โ€” Analyst, Executor, and Reviewer โ€” collaborate autonomously to diagnose the root cause, apply the fix (kubectl rollout, ServiceNow ticket update, scaling action), and verify resolution โ€” reducing Mean Time to Resolve (MTTR) by over 70%.

๐Ÿ‘ฅ

AutoGen Multi-Agent Collaboration

Three specialised agents in a GroupChat: Analyst reasons over metrics, Executor runs tools, Reviewer validates before committing changes.

๐Ÿ”ง

MCP-Connected ITSM Tools

ServiceNow, Jira, Prometheus, kubectl, and PagerDuty exposed as MCP tools โ€” the agent acts on real infrastructure with human-approval guardrails.

๐Ÿ“‰

70%+ MTTR Reduction

Automated first-response resolves known incident patterns in under 3 minutes, previously requiring 15โ€“45 minutes of on-call engineer time.

AutoGen MCP Claude Sonnet 4.5 ServiceNow Kubernetes Prometheus PagerDuty
70%โ†“
MTTR Cut
< 3 min
Auto-Resolve
3 agents
Collaborating
24 / 7
Coverage

LLMs & SLMs We Work With

We work across the full spectrum โ€” from frontier models to efficient on-device SLMs.

Proprietary

GPT-5 / GPT-4o

OpenAI

Industry benchmark for general intelligence, coding, and reasoning. Massive ecosystem with function calling, vision, and tool use support.

Coding Reasoning Vision API-Ready
128K
Context
โ˜…โ˜…โ˜…โ˜…โ˜…
Quality
$$$$
Cost
Proprietary

Claude 4.5 Sonnet

Anthropic

Exceptional at long-document analysis, nuanced writing, and complex reasoning. 200K context window ideal for enterprise document processing.

Long Docs Analysis Low Hallucination
200K
Context
โ˜…โ˜…โ˜…โ˜…โ˜…
Quality
$$$
Cost
Proprietary

Gemini 2.5 Pro

Google DeepMind

Best-in-class multimodal model with 1M+ token context. Excels at video, image, audio, and text simultaneously. Native GCP integration.

Multimodal 1M Context GCP Native
1M+
Context
โ˜…โ˜…โ˜…โ˜…โ˜…
Quality
$$$
Cost
Open Source

Llama 4 Scout / Maverick

Meta AI

Meta's flagship open-source multimodal model. MoE architecture for efficiency. Perfect for private on-premise deployments and GDPR-compliant EU workloads.

Self-Hosted MoE Arch Multimodal
128K
Context
โ˜…โ˜…โ˜…โ˜…โ˜†
Quality
Free
Cost
Open Source

Mistral Large / Mixtral

Mistral AI ๐Ÿ‡ซ๐Ÿ‡ท

European AI champion with excellent performance at lower cost. MoE architecture with strong EU regulatory alignment and data residency options.

EU Native Efficient Multilingual
32K
Context
โ˜…โ˜…โ˜…โ˜…โ˜†
Quality
$$
Cost
Open Source

DeepSeek R2 / V3

DeepSeek

Breakthrough open-source model rivalling GPT-4 at a fraction of the cost. Exceptional at mathematics and complex chain-of-thought reasoning tasks.

Open Weights Math/Code Cost Efficient
64K
Context
โ˜…โ˜…โ˜…โ˜…โ˜…
Quality
$
Cost
Proprietary

Grok 4

xAI

xAI's frontier model with real-time internet access, strong agentic capabilities, and advanced reasoning for complex multi-step workflows.

Real-time Web Agentic Reasoning
128K
Context
โ˜…โ˜…โ˜…โ˜…โ˜†
Quality
$$$
Cost
Proprietary

Command R+

Cohere

Purpose-built for enterprise RAG and tool-use. Native grounding citations, multilingual support, and ideal for Belgian multilingual (NL/FR/DE) deployments.

RAG-Native Multilingual Enterprise
128K
Context
โ˜…โ˜…โ˜…โ˜…โ˜†
Quality
$$
Cost
Open Weights

Microsoft Phi-4

Microsoft Research

14B parameter model outperforming much larger models on STEM, coding, and logical reasoning. Ideal for Azure-native and on-premise edge deployments.

14B Params Edge Ready STEM
14B
Parameters
โ˜…โ˜…โ˜…โ˜…โ˜…
Efficiency
Low
Cost
Open Weights

Phi-3.5 Mini

Microsoft

3.8B model optimised for mobile, edge, and IoT devices. Remarkable instruction-following with extremely low inference cost and latency.

3.8B Params Mobile/Edge IoT Ready
3.8B
Parameters
โ˜…โ˜…โ˜…โ˜…โ˜†
Efficiency
~Free
Cost
Open Source

Gemma 2 (2B/9B/27B)

Google DeepMind

Google's open model family with best-in-class safety. Available in 3 sizes for different compute needs โ€” from edge devices to server deployments.

Multi-size Safe AI Vertex AI
2โ€“27B
Parameters
โ˜…โ˜…โ˜…โ˜…โ˜†
Efficiency
Free
Cost
Open Source

Llama 3.2 (1B/3B)

Meta AI

Ultra-compact Meta models for on-device inference. The 3B model achieves impressive performance for classification, summarisation, and extraction at near-zero cost.

On-Device 1B/3B Params Free
1โ€“3B
Parameters
โ˜…โ˜…โ˜…โ˜†โ˜†
Efficiency
Free
Cost
Open Source

Mistral 7B / Nemo (12B)

Mistral AI ๐Ÿ‡ซ๐Ÿ‡ท

Gold standard for 7B models. Mistral Nemo offers multilingual excellence โ€” Dutch, French, German โ€” perfect for Belgian enterprise deployments.

EU Native Multilingual Self-Host
7โ€“12B
Parameters
โ˜…โ˜…โ˜…โ˜…โ˜†
Efficiency
$
Cost
Open Source

Qwen 2.5 (7Bโ€“32B)

Alibaba Cloud

Outstanding multilingual SLM family with strong coding (Qwen-Coder) and math capabilities. Multiple sizes for balancing performance and compute requirements.

Multi-size Coder Variant Multilingual
7โ€“32B
Parameters
โ˜…โ˜…โ˜…โ˜…โ˜†
Efficiency
Free
Cost
Open Source

TinyLlama / MobileLLM

Community / Meta

Sub-2B models for extreme edge deployment โ€” IoT sensors, mobile apps, browsers. TinyLlama (1.1B) runs on virtually any hardware with minimal memory.

IoT/Mobile 1.1B Params Ultra-Light
1.1B
Parameters
โ˜…โ˜…โ˜…โ˜†โ˜†
Efficiency
~Free
Cost
Open Source

StableLM Zephyr

Stability AI

3B fast-inference model optimised for edge systems and server deployments requiring low latency. Instruction-tuned variant excels at chatbot applications.

Fast Inference 3B Params Chat
3B
Parameters
โ˜…โ˜…โ˜…โ˜…โ˜†
Efficiency
Free
Cost

Tools & Frameworks We Use

Battle-tested technologies for building production-grade AI systems.

๐Ÿฆ™
Ollama
Local Inference
๐Ÿ”Œ
MCP
Tool Protocol
๐Ÿฆœ
LangChain
Orchestration
๐Ÿ•ธ๏ธ
LangGraph
AI Agents
๐Ÿค
AutoGen
Multi-Agent
๐Ÿ”
RAG Pipelines
Knowledge
๐Ÿ
Python
Core Language
๐Ÿค—
HuggingFace
Model Hub
โšก
FastAPI
API Layer
๐Ÿณ
Docker
Containers
โŽˆ
Kubernetes
Orchestration
๐Ÿ—„๏ธ
Weaviate
Vector DB
๐Ÿ“Š
MLflow
MLOps
๐Ÿ”ฅ
PyTorch
Training
๐Ÿ“ก
Terraform
IaC
๐Ÿ”
OAuth / SAML
Security

Founded in Belgium ๐Ÿ‡ง๐Ÿ‡ช

๐Ÿ“ Machelen, Flanders, Belgium

๐Ÿ‡ช๐Ÿ‡บ EU AI Act Compliant ยท GDPR Native ยท GAIA-X Aligned
๐Ÿ›๏ธ
EU HQ Brussels
๐ŸŒ
3 National Languages
๐Ÿ’ผ
Flanders Tech Hub
๐Ÿ“ก
OVHcloud EU Data

AI Innovation from the
Heart of Europe

We are a Belgium-based AI startup on a mission to make enterprise AI development accessible, GDPR-compliant, and genuinely useful. Situated in Machelen, Flanders โ€” minutes from Brussels Airport and the EU institutions โ€” we are uniquely positioned to serve European enterprise clients.

Our multilingual capabilities (Dutch, French, English, German) combined with deep AI expertise make us the ideal partner for Belgian and EU companies looking to adopt large language model technology without sacrificing data sovereignty.

๐Ÿ›ก๏ธ

GDPR & EU AI Act Compliant

All solutions built with EU data privacy regulations at the core โ€” no data leaves EU borders without consent.

๐ŸŒ

Multilingual AI (NL/FR/EN/DE)

Native support for Belgium's three official languages in all NLP and AI agent solutions we deliver.

๐Ÿš€

POC-to-Production in Weeks

Working AI prototypes in 2โ€“6 weeks leveraging cloud startup credits with minimal upfront investment.

Ready to Build Your
AI-Powered Future?

Let's discuss your AI use case. We'll help you choose the right models, architecture the right stack, and ship a working solution fast.